Remote Sensing Requirements to Drive Ecosystem Models at the Landscape and Regional Scale

To evaluate ecosystem response at the landscape and regional scale requires a search for properties that are both rich in information and measurable from space. Although it is only possible to distinguish broad surface features (forests, grasslands, wetlands, agricultural fields, urban areas) at the regional scale, present satellites can monitor seasonal and annual changes in these surface features on a daily basis with 1 km spatial resolution. Additional information on vegetation structure is desirable at the landscape scale, but satellites sensors with finer spatial resolution pass over less frequently, so a suite of satellites would be required to obtain daily coverage. At the regional scale, remote sensing can provide good estimates of incoming radiation, ambient temperature, and relative humidity. Rain radar and microwave radiometers carried on a newly launched satellite should greatly improve estimates of precipitation, but other important drivers of ecosystem processes, such as wind speed and aerosol transfers, are not yet measurable with present satellite technology. At the landscape level, improved estimates of incoming radiation can be obtained by accounting for differences in slope and aspect. Finer resolution coverage can also provide detailed information on the spatial and seasonal distribution of water in its liquid or solid state. Present remote sensing technology has the additional capability of documenting the state of vegetation phenology, a property important in all ecosystem models. To improve the modeling of ecosystem processes across complex landscapes further, we need remote sensing techniques that provide more direct measures of water stress and biochemical status of vegetation. At the same time, we need to reconfigure ecosystem models so that they can be initialized, driven, and validated more readily with present remote sensing technology